Hi Greg,
Am 03.01.2010 um 06:38 schrieb Greg Bentzinger:
So now for a brighter topic - any chance the new tool table format
might include geometry + wear? I had heard mention some changes were
planned.
The tool table definitely is a data structure which has the let's add
this and that
I'm getting ready to upgrade my machine from steppers to servos, but am
a little unsure on what would be needed.
I know I need servos. The only low cost servos I have found so far are
from keling. I'm probably going to go with the KL23-130-60 NEMA 23
sized servos. The link is here
snip
Since I assumed shareable tool information is an industry-wide problem
I had been googling around on the issue but didnt find a obvious
candidate for such a format, although a saw a manufacturer who makes
tool geometry information available in XML. Maybe there is such a
format, I just
I know I need servos. The only low cost servos I have found so far are
from keling. I'm probably going to go with the KL23-130-60 NEMA 23
sized servos. The link is here
http://www.kelinginc.net/ServoMotors.html . If anyone knows who else
sells low cost servos, I would very much be
Hi all,
I just put together a small emc2 box for my lathe-project:
http://www.anderswallin.net/2010/01/atom330itx-computer-for-emc2-lathe-control/
The performance database here
http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Latency-Test
shows a quite low jitter number of 6000 ns or so.
I was getting
The Atom330 has one CPU, if I understand correctly, but the BIOS has a
setting for HyperThreading, so Ubuntu thinks there are two CPUs. Has
anyone experimented with HT on/off to see the effect on realtime
latency? (would I need an SMP-realtime kernel?)
The Atom330 is a dual-core CPU. That
Le mardi 5 janvier 2010 11:29:52, Anders Wallin a écrit :
I was getting double that, or 12-16000 ns when running glxgears,
firefox, etc. together with latency-test.
not so bad... this is what I get with my PIII box :)
The Atom330 has one CPU, if I understand correctly, but the BIOS has a
The Atom330 is a dual-core CPU.
interesting, the default Real-Time kernel apparently does not see the
second core at all?
Since when I look at the resource-monitor there is only one graph for cpu load.
So with HyperThreading the Atom330 looks like 4 cores to the OS?
Question: do you need
2010/1/5 Anders Wallin anders.e.e.wal...@gmail.com:
interesting, the default Real-Time kernel apparently does not see the
second core at all?
Since when I look at the resource-monitor there is only one graph for cpu
load.
That is normal if you use the ISOLCPUS option, the core dedicated to
Andy Pugh wrote:
2010/1/5 Anders Wallinanders.e.e.wal...@gmail.com:
interesting, the default Real-Time kernel apparently does not see the
second core at all?
Since when I look at the resource-monitor there is only one graph for cpu
load.
That is normal if you use the ISOLCPUS
2010/1/5 Stephen Wille Padnos spad...@sover.net:
That is normal if you use the ISOLCPUS option, the core dedicated to
the RT threads doesn't show up in the resource monitor.
Actually, that's not true. The core is still there, and is still
monitored by the normal tools. The only thing
Andy Pugh wrote:
2010/1/5 Stephen Wille Padnosspad...@sover.net:
That is normal if you use the ISOLCPUS option, the core dedicated to
the RT threads doesn't show up in the resource monitor.
Actually, that's not true. The core is still there, and is still
monitored by the normal
2010/1/5 Andy Pugh a...@andypugh.fsnet.co.uk:
What I was doing was repeating section 7.2 from that link,
And I need to learn to read. Sorry.
For some reason I thought that Anders was talking about the SMP kernel
that Alex mentioned. There was a whole hour between the posts, plenty
of time for
Anders Wallin wrote:
Brushless is not much more expensive so you could look at those too.
I don't know much about servos at all, but I was under the impression
that brushless servos had less angular position accuracy then brushed dc
servos? It looked like the brushless dc motors had more
On 1/5/2010 11:49 AM, Flying Electron wrote:
Anders Wallin wrote:
Brushless is not much more expensive so you could look at those too.
I don't know much about servos at all, but I was under the impression
that brushless servos had less angular position accuracy then brushed dc
Dave wrote:
On 1/5/2010 11:49 AM, Flying Electron wrote:
Since you have a mesa card, maybe you or anyone else that has a mesa
card can explain how the servo PID works with the mesa in EMC. Does EMC
do the PID calculations and just send a command to the mesa board or
does the mesa board
Alex,
I glossed over your references and am shocked, it's just a single
*table* ;-)
I'm proposing a much more moderate goal: to encode the current EMC
tool table such that it can be reused by other programs, and with some
self-description features (field names, types, version number,
Brushless servos have a permanent magnet rotor.The coils that make
up the stator ( the stationary coils) are up against the outside walls
of the motor so any heat from the coils can be transferred out of the
motor rather easily.On a brush servo motor, the rotor has a wound
rotor
Flying Electron wrote:
Between your information and a post I found on CNC zone at
http://www.cnczone.com/forums/showthread.php?t=91306 I think I got a
handle on how things are supposed to work together.
Is the mesa 7I43 card the recommended card for EMC out of the mesa products?
It's
Flying Electron wrote:
Is the mesa 7I43 card the recommended card for EMC out of the mesa products?
The 7i43 works well if you have a good EPP parport (ie, not a NetMos
chip, and not a USB parallel port), and it's inexpensive. It's not my
favourite because the parallel port is kind of slow.
Michael Haberler wrote:
Alex,
I glossed over your references and am shocked, it's just a single
*table* ;-)
I'm proposing a much more moderate goal: to encode the current EMC
tool table such that it can be reused by other programs, and with some
self-description features (field names,
On 1/5/2010 12:15 PM, Sebastian Kuzminsky wrote:
Dave wrote:
On 1/5/2010 11:49 AM, Flying Electron wrote:
Since you have a mesa card, maybe you or anyone else that has a mesa
card can explain how the servo PID works with the mesa in EMC. Does EMC
do the PID calculations and just
FBICS has a tool catalog with many features. The data is stored in a STEP
Part 21 file.
IMO, it would be worthwhile to examine the features Tom envisioned even if
Part 21 won't be used.
I think there is at least some discussion of the catalog in the FBICS docs
on Tom Kramer's NIST page. I could
One other thing that would be nice in the tool table would be to specify the
units of measurement. Then if a tool was metric or inch there would be no need
for manual unit conversion before entering the value.
John Figie
RT tasks don't show up on multiprocessor or uniprocessor systems, since
they run in the kernel. (or I think that's the reason anyway) It
Actually they run in another domain.
IPIPE (or ADEOS previously) is the way that more domains co-exist on the
same hardware.
The RTAI domain is a separate
On Tue, Jan 05, 2010 at 01:00:39PM -0500, Stephen Wille Padnos wrote:
So yes, making changes so that extra information could be added as
necessary would be great.
Within the existing plain text format, that would be nifty indeed.
Making a format that humans can't read or write without
On Wed, Jan 06, 2010 at 02:59:13PM +1100, Erik Christiansen wrote:
If we don't watch out, there is a risk that the only way to read or edit
our configuration files would be some GUI monstrosity! (Hopefully, I
exaggerate horribly. :-)
I sure sympathize with this worry. XML is very hostile to
On Tue, Jan 05, 2010 at 10:18:59PM -0600, Chris Radek wrote:
Here is what the new tool table format looks like:
http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=configs/sim/sim.tbl;hb=tlo_all_axes
Now that looks very business-like.
You can see tool number, pocket, diameter, Z offset,
Just poking in here, but why not use comma delineators. Then you can by
default have, say, 16 parameters per tool, most unused for now, but allows
expansion of parameters at any stage. So each line would look something
like;
1, T1, P1, D0.125000, Z+0.511000 , , , , , , , , , , , , ;1/8 end
On Wed, Jan 06, 2010 at 04:21:33PM +1100, Erik Christiansen wrote:
I'll figure out how to use git, and take a peek at his branch.
My git clone looks like it'll take 2.8 hours, at 8 KiB/s. That doesn't
seem normal, with the wiki saying the 80 MB downloads in 3 minutes on a
3 megabit connection.
30 matches
Mail list logo
